The science resulting from our robot exploration is astounding. Nearly all photos of outer space objects beyond the earth and moon are robot driven, and the high quality photos of Jupiter, etc are entirely robot driven. Compare that to a Mars rock.
BTW, we can't send a human to Mars right now. It's technically impossible- well, I mean, we could deliver their dead body to the surface, but that's not super useful.
A single geologist on Mars would constantly have to check back in on their plans.
Mars is not. It's a minimum of a 1-hour round-trip. So every tiny little rover movement has to be planned in advance, unless you can figure out how to automate it, which only goes so far: you're not going to put an AI with a geologist's education in a rover.
This is not likely to ever change, unless you can 1) figure out how to make FTL communications work (subspace?), or 2) figure out how to make AIs with human-level intelligence (at which point we humans would be completely obsolete and likely slated for extermination). Good luck with either of those. Until then, sending actual humans is the only way to get a lot done in a reasonable amount of time.
The nice photos you talk about from robotic exploration is the result of completely pre-planned missions, with very little possibility of needing to make fast decisions on-site. When you're just flying around the Jupiter system or flying through the Pluto system, that's really not that hard; the planet and moons' locations can be predicted with a very high degree of accuracy, and the data they're gathering is all planned far in advance. It's not like walking around a planet's surface and deciding which rock to investigate, or which rock to ignore, or what a good place to dig would be. You can do all that by remote control, but it's a whole lot slower because you need to have humans verify everything, so what would take you an hour with a human on-site now takes you days.
The beauty of large-scale robotic exploration is that you can use all the analysts on earth to analyze the results. Compared to a single human.
Yes, NASA has made mission changes post-launch, but these are usually simple changes, consisting basically of single orders to redirect the craft along a new trajectory. A human walking around looking at rocks makes those decisions on a second-by-second basis.
Growing crops in space has no real scientific benefits for the nearterm future. It's a "nice to have" not a "visionary" thing.